Evaluating the interpretability of a hierarchical fuzzy rule-based model for shipbreaking

Abstract

Machine learning models can provide valuable decision support in many real-world applications. However, a model must be interpretable to those using it. This paper explores the use of post-hoc model interpretability methods in combination with an intrinsically interpretable model design to create a model that is interpretable to both a model designer and a model end user. A hierarchical fuzzy rule-based model is trained with a genetic algorithm on a real-world shipbreaking use case and the performance-interpretability trade-off of the model with respect to a random forest model is discussed. Further, an interesting pattern was found using the post-hoc interpretability method SHapley Additive exPlanations (SHAP), with potential implications for the future design of hierarchical fuzzy rule-based models.

Keywords

Genetic fuzzy rule-based model / fuzzy logic / interpretable machine learning / artificial intelligence

References

[1] European Union. Proposal for a regulation of the European Parliament and of the council laying down harmonised rules on artificial intelligence (artificial intelligence act) and amending certain union legislative acts; 2021. Available from: https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=celex:52021PC0206#document2 [Last accessed on 31 Oct 2025].

[2] Rudin C..Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead.Nat. Mach. Intell.2019;1:206-15 PMCID:PMC9122117

[3] Zadeh L. A..Fuzzy sets.Inf. Control.1965;6:338-53

[4] Alonso J. M..Special issue on interpretable fuzzy systems.Inf. Sci.2011;10:4331-39

[5] Alonso Moral, J. M.; Castiello, C.; Magdalena, L.; Mencar, C. Designing interpretable fuzzy systems. In: Explainable fuzzy systems, studies in computational intelligence. Cham: Springer International Publishing; 2021. pp. 119-68.

[6] Alonso Moral, J. M.; Castiello, C.; Magdalena, L.; Mencar, C. Design and validation of an explainable fuzzy beer style classifier. In: Explainable fuzzy systems, studies in computational intelligence. Cham: Springer International Publishing; 2021. pp. 169-217.

[7] Wang L. X..Analysis and design of hierarchical fuzzy systems.IEEE. Trans. Fuzzy. Syst.1999;7:617-24

[8] Zhang Y.,Wang S..Deep Takagi–Sugeno–Kang fuzzy classifier with shared linguistic fuzzy rules.IEEE. Trans. Fuzzy. Syst.2018;26:1535-49

[9] Razak, T. R.; Garibaldi, J. M.; Wagner, C.; Pourabdollah, A.; Soria, D. Interpretability indices for hierarchical fuzzy systems. In: 2017 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE); 2017. pp. 1-6.

[10] Magdalena, L. Designing interpretable hierarchical fuzzy systems. In: 2018 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE); 2018. pp. 1-8.

[11] Alonso, J. M.; Cordon, O.; Quirin, A.; Magdalena, L. Analyzing interpretability of fuzzy rule-based systems by means of fuzzy inference-grams; 2011, pp. 181-85. Available from: https://sci2s.ugr.es/sites/default/files/ficherosPublicaciones/1394_Alonso-etal-WConSC11.pdf [Last accessed on 31 Oct 2025].

[12] Kokkotis C.,Moustakidis S.,Tsaopoulos D..Explainable machine learning for knee osteoarthritis diagnosis based on a novel fuzzy feature selection methodology.Phys. Eng. Sci. Med.2022;45:219-29 PMCID:PMC8802106

[13] Li S.,Feng K..Composite neuro-fuzzy system-guided cross-modal zero-sample diagnostic framework using multisource heterogeneous noncontact sensing data.IEEE. Trans. Fuzzy. Syst.2025;33:302-13

[14] Cai X.,Ning Z.,Chen J..A many-objective multistage optimization-based fuzzy decision-making model for coal production prediction.IEEE. Trans. Fuzzy. Syst.2021;29:3665-75

[15] Deveci M.,Karagoz S..An interval type-2 fuzzy sets based Delphi approach to evaluate site selection indicators of sustainable vehicle shredding facilities.Appl. Soft. Comput.2022;118:108465

[16] Huang W.,Chen L.,Cao W..Multi-objective drilling trajectory optimization using decomposition method with minimum fuzzy entropy-based comprehensive evaluation.Appl. Soft. Comput.2021;107:107392

[17] Pickering L.,De Baets B..A narrative review on the interpretability of fuzzy rule-based models from a modern interpretable machine learning perspective.Int. J. Fuzzy. Syst.2025;

[18] Ministry of Infrastructure and Water Management. About the ILT; 2023. Available from: https://english.ilent.nl/about-the-ilt [Last accessed on 31 Oct 2025].

[19] European Union. Regulation (EU) No 1257/2013 of the European Parliament and of the council of 20 November 2013 on ship recycling and amending regulation (EC) No 1013/2006 and directive 2009/16/EC text with EEA relevance; 2013. Available from: https://eur-lex.europa.eu/eli/reg/2013/1257/oj/eng [Last accessed on 31 Oct 2025].

[20] Barua S.,Hossain M. M..Environmental hazards associated with open-beach breaking of end-of-life ships: a review.Environ. Sci. Pollu. Res.2018;25:30880-93

[21] The Kingdom of the Netherlands. Staatsblad van het Koninkrijk der Nederlanden; 2021. Available from: https://zoek.officielebekendmakingen.nl/stb-2021-499.html [Last accessed on 31 Oct 2025]

[22] Hadwick, D.; Lan, S. Lessons to be learned from the dutch childcare allowance scandal: a comparative review of algorithmic governance by tax administrations in the Netherlands, France and Germany. WTJ 2021, 13, 609-45. Available from: https://ssrn.com/abstract=4282704 [Last accessed on 31 Oct 2025].

[23] Kazim, E.; Koshiyama, A. Explaining decisions made with AI: a review of the co-badged guidance by the ICO and the Turing Institute; 2020.

[24] Carvalho D. V.,Cardoso J. S..Machine learning interpretability: a survey on methods and metrics.Electronics2019;8:832

[25] Linardatos P.,Kotsiantis S..Explainable AI: a review of machine learning interpretability methods.Entropy2020;23:18 PMCID:PMC7824368

[26] Lundberg, S. M.; Lee, S. I. A unified approach to interpreting model predictions; 2017. Available from: https://proceedings.neurips.cc/paper_files/paper/2017/file/8a20a8621978632d76c43dfd28b67767-Paper.pdf [Last accessed on 31 Oct 2025].

[27] Scikit-Learn. Permutation importance vs random forest feature importance (MDI); 2023. Available from: https://scikit-learn.org/stable/auto_examples/inspection/plot_permutation_importance.html [Last accessed on 31 Oct 2025].

[28] Gini, C. Variabilità e mutabilità: contributo allo studio delle distribuzioni e delle relazioni statistiche. [Fasc. I. ]. Studi economico-giuridici pubblicati per cura della facoltà di Giurisprudenza della R. Università di Cagliari. Tipogr. di P. Cuppini; 1912. Available from: https://books.google.se/books?id=fqjaBPMxB9kC [Last accessed on 11 Nov 2025].

[29] Breiman L..Random forests.Mach. Learn.2001;45:5-32

[30] Shapley, L. A value for n-Person games. In: Kuhn, H.; Tucker, A.; editors. Contributions to the theory of games II. Princeton University Press; 1953, pp. 307-17.

[31] Lundberg S. M.,Chen H..Explainable AI for trees: from local explanations to global understanding.CoRR2019;1905.04610

[32] NGO Shipbreaking Platform. Annual lists of scrapped ships; 2023. Available from: https://shipbreakingplatform.org/annual-lists/ [Last accessed on 31 Oct 2025].

[33] International Maritime Organization. GISIS; 2023. Available from: https://gisis.imo.org/ [Last accessed on 31 Oct 2025].

[34] European Maritime Safety Agency. THETIS-EU; 2023. Available from: https://portal.emsa.europa.eu/web/thetis-eu/ [Last accessed on 31 Oct 2025].

[35] Fernandes E. R. Q.,Yao X..Ensemble of classifiers based on multiobjective genetic sampling for imbalanced data.IEEE. Trans. Knowl. Data. Eng.2020;32:1104-15

[36] Mohammed, R.; Rawashdeh, J.; Abdullah, M. Machine learning with oversampling and undersampling techniques: overview study and experimental results. In: 2020 11th International Conference on Information and Communication Systems (ICICS); 2020. pp. 243-48.

[37] Ruspini E. H..A new approach to clustering.Inf. Control.1969;15:22-32

[38] Assilian, S. Artificial intelligence in control of real dynamic systems. Queen Mary University of London; 1974. Available from: http://qmro.qmul.ac.uk/xmlui/handle/123456789/1450 [Last accessed on 31 Oct 2025].

[39] Mamdani E. H..Application of fuzzy algorithms for control of simple dynamic plant.Proc. Inst. Elect. Eng.1974;121:1585-88

[40] Kruse R.,Gebhardt J..Foundations of fuzzy systems. Chichester, West Sussex, England, New York: Wiley & Sons; 1994.

[41] Pickering, L.; Cohen, K. Toward explainable AI - genetic fuzzy systems - a use case. In: Rayz, J.; Raskin, V.; Dick, S.; Kreinovich, V.; editors. Explainable AI and other applications of fuzzy techniques. Cham: Springer International Publishing; 2022. pp. 343-54.

[42] Golberg D..Genetic algorithms and machine learning.Mach. Learn.1988;3:95-9

[43] Holland J. H..Genetic algorithms.Sci. Am.1992;267:66-73

[44] Mirjalili, S. Evolutionary algorithms and neural networks. In: Studies in computational intelligence. Springer; 2019, pp. 1-170.

[45] Kochenderfer, M. J.; Wheeler, T. A. Algorithms for optimization. MIT Press; 2019. Available from: https://algorithmsbook.com/optimization/files/optimization.pdf [Last accessed on 31 Oct 2025].

[46] Saito T..The precision-recall plot is more informative than the ROC plot when evaluating binary classifiers on imbalanced datasets.PLoS One2015;10:e0118432 PMCID:PMC4349800

[47] Du M.,Hu X..Techniques for interpretable machine learning.Commun. ACM2019;63:68-77

[48] Pedregosa F.,Gramfort A..Scikit-learn: machine larning in Python.J. Mach. Learn. Res.2011;12:2825-30

[49] Breiman L.,Olshen R. A..Classification and regression trees. Belmont, CA: Wadsworth International Group; 1984.

[50] Plonski, P. Extract rules from decision tree in 3 ways with Scikit-Learn and python. MLJAR; 2021. Available from: https://mljar.com/blog/extract-rules-decision-tree/ [Last accessed on 31 Oct 2025].

[51] Wang H.,Hancock J. T..Feature selection strategies: a comparative analysis of SHAP-value and importance-based methods.J. Big. Data.2024;11:1-16

[52] Alcala-Fdez J.,Herrera F..A fuzzy association rule-based classification model for high-dimensional problems with genetic rule selection and lateral tuning.IEEE. Trans. Fuzzy. Syst.2011;19:857-72